1. Field of the Invention
The present invention relates to a technique for appropriately correcting a color misregistration.
2. Description of the Related Art
Conventionally, an imaging apparatus such as a digital camera is used for a variety of applications. However, various aberrations which an imaging lens used for forming an object image has, become a cause for deteriorating an image quality of the object image. For example, lateral chromatic aberration becomes a cause for a color misregistration on the formed object image. Since the number of pixels of an image sensor used in the imaging apparatus is increasing year after year, a unit pixel size is showing a tendency to decrease, and even such a degree of lateral chromatic aberration that has conventionally given little problems has become a main cause for deteriorating the image quality.
As a technique for correcting such a color misregistration by image processing, a technique for acquiring a color misregistration amount to be corrected (that is, a correction amount) from image data of a correction target is discussed. As the technique for acquiring a color misregistration amount from image data, there is a method in which a correlation between respective color components at edge portions of the image data is utilized.
For example, Japanese Patent Application Laid-Open No. 2000-299874 discusses a method for acquiring a color misregistration amount by searching for a position at which a total of differences of distances of pixels of respective color components corresponding to edges becomes minimal. In addition, Japanese Patent Application Laid-Open No. 2006-20275 discusses a method for detecting edges from raw data and determining a displacement width, in which an error of array of two types of color components at edge portions becomes minimal, as a color misregistration width.
However, the method for acquiring the amount of color misregistration from image data as described above may not allow an appropriate amount of color misregistration to be acquired in some cases, with respect to image data containing a lot of noises, an object tinted with a color like lateral chromatic aberration at edges, or image data which has too low a contrast to detect the edges.
Although there is also a method for determining a correction amount based on lens design data, it is conceivable that an accurate correction amount of lateral chromatic aberration cannot be obtained depending on a manufacturing error of a lens or shooting conditions such as color temperature of a light source or an object.